Abstract
BACKGROUND: Glioma is the most common malignant tumor of the brain. The existence of metastatic tumor cells is an important cause of recurrence even after radical glioma resection. METHODS: Single-cell sequencing data and high-throughput data were downloaded from GEO database and TCGA/CGGA database. By means of PCA and tSNE clustering methods, metastasis-associated genes in glioma were identified. GSEA explored possible biological functions that these metastasis-associated genes may participate in. Univariate and multivariate Cox regression were used to construct a prognostic model. RESULTS: Glioma metastatic cells and metastasis-associated genes were identified. The prognostic model based on metastasis-associated genes had good sensitivity and specificity for the prognosis of glioma. These genes may be involved in signal pathways such as cellular protein catabolic process, p53 signaling pathway, transcriptional misregulation in cancer and JAK-STAT signaling pathway. CONCLUSION: This study explored glioma metastasis-associated genes through single-cell sequencing data mining, and aimed to identify prognostic metastasis-associated signatures for glioma and may provide potential targets for further cancer research.